© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 129547
    Prev
    Next
    INDEX
    Explanations

    The neuron fires on degree-of‐intensity markers—e.g. explicit numerical/percentage values and strong intensifier adverbs (completely, totally, unacceptably).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    absence of perfection

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    anhas
    -1.12
    ท่าน
    -1.10
     pratiquement
    -1.09
    étant
    -1.08
    paravant
    -1.05
     qualquer
    -1.02
     تقری
    -1.02
    amiento
    -1.01
     instead
    -1.00
    没有任何
    -1.00
    POSITIVE LOGITS
     there
    1.32
     you
    1.22
    </h2>
    1.16
     because
    1.15
     Even
    1.14
     but
    1.11
     what
    1.02
     even
    1.02
     nor
    1.01
     sondern
    1.01
    Activations Density 0.036%

    No Known Activations